Figure 7.

CAV1 secretion from LNCaP cells is mediated by an autophagy-dependent pathway. (A) Representative confocal microscopy images from three independent experiments showing co-localization between YFP-CAV1 and LC3B in LNCaP cells with or without 6-h serum starvation. Endogenous LC3B was immuno-stained with rabbit anti-LC3B primary antibodies followed by a secondary Alexa555 fluorescent labeling. Knockdown of ATG5, ATG9A or ATG12 and 3-MA (5 mM, 6 h) treatment rescued YFP-CAV1 expression and inhibited autophagosome formation. Single channel images were converted to black and white and the contrast was inverted. The enlarged region demonstrates the overlapping distribution between YFP-CAV1 and LC3B. Scale bar: 10 τm. (B) Quantification of the number of LC3B vesicles in the confocal images. One-way ANOVA was performed for statistical analysis from three experiments, n ≥ 50 cells for each experiment. (siScrambled fed vs. siScrambled starved: p = 0.0003; siATG5 starved vs. siScrambled starved: p = 0.0013; siATG9A starved vs. siScrambled starved: p = 0.0022; siATG12 starved vs. siScrambled starved: p = 0.0005; 3-MA starved vs. siScrambled starved: p = 0.0014). (C) Representative western blots showing YFP-CAV1 protein levels in WCL and S100 fractions of LNCaP cells equivalently transfected with YFP-CAV1 (10 τg DNA per 150 × 25 mm dish). The knockdown efficiency was assessed by the detection of ATG5, ATG9A and ATG12 protein levels. (D) The ratio of LC3B-II:LC3B-I (LC3B-II:I) was calculated as an indicator of autophagic levels in each group. A significant increase in the LC3B-II:I ratio of the serum-starved siScrambled cells (2.8) compared to the fed siScrambled cells (1.3) indicates the induction of autophagy. One-way ANOVA was performed for statistical analysis from three experiments. (siScrambled fed vs. siScrambled starved: p = 0.0262; siATG5 starved vs. siScrambled starved: p = 0.0307; siATG9A starved vs. siScrambled starved: p = 0.0497; siATG12 starved vs. siScrambled starved: p = 0.0223; 3-MA starved vs. siScrambled starved: p = 0.0137). (E) Densitometry analysis of western blots shown in (D) demonstrate that the co-transfection with ATG5, ATG9A or ATG12 siRNAs and 3-MA treatment significantly (siATG5 starved vs. siScrambled starved: p = 0.0047; siATG9A starved vs. siScrambled starved: p < 0.0001; siATG12 starved vs. siScrambled starved: p = 0.0031; 3-MA starved vs. siScrambled starved: p < 0.0001) rescued the downregulation of YFP-CAV1 (siScrambled fed vs. siScrambled starved: p = 0.0246) in the WCL of serum-starved cells compared to the fed cells. One-way ANOVA was performed for statistical analysis. (F) The secretion (%) of YFP-CAV1 into the S100 fraction from LNCaP cells was significantly upregulated (siScrambled fed vs. siScrambled starved: p = 0.0011) upon serum starvation. The blockage of serum starvation-induced autophagy via ATG5, ATG9A, ATG12 knockdowns or 3-MA treatment downregulated the secretion levels of YFP-CAV1. Statistically-significant effects (one-way ANOVA) were observed in ATG9A siRNA (siATG9A starved vs. siScrambled starved: p = 0.0152), ATG12 siRNA (siATG12 starved vs. siScrambled starved: p = 0.0056) and 3-MA (3-MA starved vs. siScrambled starved: p = 0.0007) treated groups